Comment on page
It allows you to train the NLP engine with your FAQs so that AI can understand user's queries and answer them automatically.
Please read the introduction in order to understand how FAQ works.
In the previous articles you have learned about how to create conversational flows and deploy them. There are some users who do not select given options from the flow but type some other query they might be having. For example chatbot flow is asking them to know more about product features but user is interested in pricing. So user might type a query like "how much it would cost" or "What are the charges" etc. To automate responses of such queries, you can train the NLP engine with the sample FAQ queries. After that NLP engine will try to match the similarity of query with the training data you have provided (It does not do the exact matching). If it finds user's query closer to the training data, it would respond with the configured response automatically. Otherwise it will trigger the Transfer playbook to transfer the chat to the human.
- 1.Go to AI Builder from the left Menu
- 2.Go to FAQs.
- 3.In the FAQs section you can see the currently trained queries.
- 5.In the left section "if user says something similar to" enter a phrase or query user might have. Press enter to add it to the list. Similarly add multiple variations of the same query. The more variations you add, better it would work but be sure to add phrases similar to user queries without any spelling mistakes, rather than entering only keywords.
- 6.In the right section you can provide responses to the above queries. It is discussed in the separate section below.
There are 3 type of responses.
When instead of giving a plain text response, you wish to engage the end user with a pre-trained bots that come with the platform, choose the Pre-trained Bot Response & select the bot. Below is the screenshot of FAQ configuration with Pre-trained Bot response.
Custom Bot Response
When you wish to hook the end-user with a bot that you have created, you can use the custom bot response.
Below is the screenshot of FAQ configuration with Custom Bot response.